Available online at www.sciencedirect.

com

ScienceDirect
Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163

Social and Behavioral Sciences Symposium, 4th International Science, Social Science,
Engineering and Energy Conference 2012 (I-SEEC 2012)

Failure Analysis of a Two High Gearbox Shaft

Charnont Moolwana,*, Samroeng Netpub
a
Department of Industrial Engineering, Faculty of Engineering, Kasem Bundit University, Bangkok, 10250, Thailand
b
S.H.K. Engineering Co., Ltd. Samutprakarn, 10130, Thailand

Abstract

This paper reports the results of failure analysis of a two high gearbox shaft of a gearbox in a hot steel rolling mill in
Thailand which fail prematurely after about 15,000 hours of service. Standard procedures for failure analysis were employed
in this investigation. The results showed that the shaft failed by fatigue fracture. Beach marks on the fracture surface were
clearly visible. Fatigue cracks were initiated at the corners of the wobbler. Relatively small final fracture area of the fracture
surface indicated that the shaft was under a low stress at the time of failure.
It is concluded that the shaft failed by fatigue fracture and that premature failure occurred due to high stress concentration at
the corners of the wobbler of the shaft which led eventually to fatigue crack initiation, crack growth, and final fracture.
Improved design and machining practice suggested that this would help prolong service life of the component.

©
© 2013 TheAuthors.
2013 The Authors. Published
Published by Elsevier
by Elsevier Ltd. access under CC BY-NC-ND license.
Ltd. Open
Selection and/orpeer-review
Selection and/or peer-review under
under responsibility
responsibility of Faculty
of Department of Science
of Planning andand Technology,
Development, Kasem
Kasem Bundit
Bundit University,
University Bangkok.
- Bangkok

Keywords:

1. Introduction

Shafts are extensively used in machines and numerous engineering components including gearboxes. Failures
of shafts not only result in replacement cost, but also in process downtime. This could have a drastic effect on
productivity and, more importantly, late delivery. In the case being investigated, for example, the downtime was 3
days, and 1,800 metric tons of steels were lost before the failed shaft could be replaced.
Shaft failures may result from many causes including faulty designs, improper applications or manufacturing
errors. Design errors include such things as improper gear geometry, wrong materials, poor materials quality,
inappropriate lubrication system, and several others. Application errors include things such as improper mounting
and installation, poor cooling, inadequate lubrication, and poor maintenance. Manufacturing errors could be poor
machining or faulty heat treatments [1]. A gear shaft is usually subject to a high torsion and a bending moment as

* Corresponding author. E-mail address: charnont_99@hotmail.co.th

1877-0428 © 2013 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Selection and/or peer-review under responsibility of Department of Planning and Development, Kasem Bundit University - Bangkok
doi:10.1016/j.sbspro.2013.08.491
 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163 155

well as cyclic stresses which, when combined, may cause fatigue in the shaft. These factors may be influenced by
stress concentrations which decrease the fatigue life and lead to the fracture of gear shaft. The fatigue fracture is
one of the most common causes of shaft failure. Fatigue failures are insidious and are therefore important
considerations in mechanical designs [2]. Despite preventive measures taken during the design stage, fatigue
failure can still occur due to either defect introduced during fabrication and/or degradation of shafts during
service [3]. Gear shafts are usually subject to a combination of high torsional loads and bending moments in
cyclical manners which lead to fatigue cracks in the shafts [4]. The common failure modes of gear shafts were
failed to be in decreasing order of frequency, as follows: fatigue, impact fracture, wear, and stress rupture [5].
The fatigue crack growth rate of the heat-affect zone (HAZ) from pre-heat is lower than that of post-weld heat
treatment, indicated that the lowest crack growth rate and best fatigue resistance result from pre-heat treatment at
350 degrees without post-weld heat treatment. When steels are welded, the optimized base metal properties are
altered by the localized weld thermal cycles. The result is the creation of non-equilibrium microstructures in the
weld fusion zone (FS) and heat affected zone (HAZ).
This paper aims at identifying the cause of failure of a 2 high gearbox shaft so that the reoccurrence of similar
failure can be avoided in the future.

2. Background

The failed shaft being investigated was a bottom shaft used in two high gearbox in a continuous hot rolling
steel rebars mill in Thailand. The mill produced steel rebars 12 to 20 mm in diameter with a capacity of 35 tonnes
per hour. The mill was designed for rolling steel billets with cross-sectional area of 120 mm x 120 mm square and
3 meters long. The gearbox was installed at pass number two and was driven by an electric AC motor of 400 kW.
The two high gearbox has two output shafts as shown in Fig. 2, and both the rotational speed of shafts was 19
rpm.
Rolls and two high gearbox shafts were connected by steel spindles, cast iron couplings and wooden blocks
prevented displacement of coupling in the wobbler grooves of the spindle during operation of the mill as shown in
Fig 1. Due to this design of the spindle connection, the substantial weight of the parts (about 450 kg) was
unbalanced.
The shaft failed prematurely after about 15,000 hours of service. Normally, a gearbox has an expected working
life of around 30,000-50,000 hours [6].
The two high gearbox was driven rolls by details of the 2-hi gearbox assembly, the location of the shaft fracture
and key dimensions are shown in Fig. 2.

2 high gearbox Coupling

Failed shaft Spindle

Fig. 1 Spindle connection between rolls and 2 high gearbox

156 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163

a) Shaft with herringbone gear. b) Key dimension of the failed shaft

Fig. 2 Schematic diagram of 2 high gearbox shaft

3. Investigation procedure

The failed shaft was first inspected visually and macroscopically. Relevant dimensions were measured and
details of operating conditions noted.
There are two hardness measurement regions. The first start measured from welding surface (A). The second
start measured from outside of the wobbler to the interior. It was carried out on a polished specimen using a
Mitutoyo model HV-115 Vickers hardness tester with a load of 300g.
Chemical analysis of the material was performed in order to identify the type of steel used. Chemical
composition was analysed using an optical emission spectrometer (Spectrolab Model Dv-4/202472-778A) .
Standard metallographic sp ecimen preparation procedur es were empl oyed. Microstructure of the
specimen was examined under an optical microscope (LECO: IA32-Image analysis system).
Load calculation using the data from actual operating conditions, electrical power (W p) was calculated using
equation 1 [12]. Transmitted torque (T) was calculated using equation 2 [13].

Wp V A 3 cos (1)

60,000W p
T (2)
3.14 x 2 xN

where WP is electrical power in kW. V is voltage (V) which is 380 volts. A is ampere (Amp) which is 600 Amp.
cos is power factor which is 0.8. T is the transmitted torque in N.m. And N is the speed of the shaft which is
19.0 rpm.
Tangential and radial loads on the pitch were calculated using equations 3 and 4 [9].
T
Wt (3)
D/2
Wr Wt tan (4)
 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163 157

where Wt is the tangential load in MPa. Wr is the radial load in MPa. D is pitch diameter which is 435 mm in
this case. is pressure angle of the gear which is 20 degrees.
Inclusions were analyzed using Energy Dispersive Spectroscopy technique. The scanning electron microscope
employed was JEOL Model JSM-6380 LV with OXFORD IE 300 Energy Dispersive Spectroscopy (EDS)
microanalysis system. The modes of analysis used in this investigation are; point analysis, line scanning, and
mapping.
An FE model for stress analysis of the gear shaft was performed by ANSYS Mechanical software. The
geometry of the model reflects the actual dimensions of the shaft. All of the solid elements are defined using the
207 GPa and the Poission ratio of 0.3.

4. Results

4.1 Visual examinations

The failed shaft was taken out from the two high gearbox and examined visually. The fracture occurred at
wobbler of bottom shaft. A general appearance of the failed shaft is as shown in Fig. 3a. The wobbler surface on
working site of the shaft was heavy wear only in some area as shown in Fig. 3b indicated impact load and small
area contact between the wobbler and coupling. And weld overlay feature surface on the wobbler can be seen as
shown in Fig. 3b.

a) the fracture point of the shaft b) the working site of the shaft

Fig. 3 The fracture surfaces of the failed shaft

4.2 Fracture surface examination

The fracture surfaces were slightly curved; i.e., on convex and the other concave. The fracture surface has two
distinct portions both of which are rather flat as shown in Fig. 4. The final fracture is characteristic of a high cycle
fatigue. The final fracture surface area is small, approximately 10% of the total fracture surface as shown in Fig 4
indicating that the shaft material was adequate for the low applied stresses.
Beach marks are clearly visible on fracture surface as shown in Fig. 5, characteristics of fatigue and the
propagation from the initiation site, and the location of final fracture. The fatigue crack initiated at corners of the
wobbler as shown in Fig. 5b.
158 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163

(a) (b)

Fig. 4 The fracture surfaces of the failed shaft. (a) A side, (b) B side of Fig. 3a

a) crack initiation sites. b) beach marks on the fracture surface.

Fig. 5 The fracture surfaces of the failed shaft

4.3 Hardness measurements

There are two microhardness values measured at various distance from the initial crack region (start from point
A) and measured at various from outside (start from B) to toward of the shaft. The hardness values were found to
be maximum 290 HV and minimum 210 HV. The hardness values measured at various distances from outside
(started from B) toward the shaft are as shown in Fig .6. The hardness values were found to be between 210- 240
HV which was within specified limits of commercial machinery steels [10]. The hardness readings indicated a
relatively high hardness in the HAZ on the shaft. The microhardness measurement (start from surface A) revealed
a hard surface that gets progressively softer towards the core as shown in Fig. 6.
 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163 159

a) hardness testing of the sample b) distance from surface of the failed shaft (mm)

Fig. 6 Hardness of the failed shaft

4.4 Chemical composition analysis

The average values of the chemical composition of shaft material are shown in Table 1. The ranges for the
composition of AISI 1045 steel are also included in Table 1. The compositions of the shaft material revealed it to
be medium carbon low alloy steel to AISI 1045 standard [14]. The AISI 1045 belongs to a class of high strength
steels. AISI 1045 is a low alloy steel suitable for most engineering and machinery components.

Table 1 Chemical composition of the failed shaft and AISI 1045 (%wt)

Material C Si Mn Ni Cr Mo P S
Failed Shaft 0.450 0.281 0.634 0.215 0.205 0.037 0.019 0.027
AISI 1045 0.38-0.45 0.4 (Max) 0.60-0.90 0.30

4.5 Microstructure examination

The microstructure of the failed shaft on crack origin, unaffected region from weld and core region that were
not tempered, are shown in Fig. 7. The microstructure of core was a mixture of pearlite and ferrite as shown in
Fig.7(a). The structure of the core is a typical structure for low alloy medium carbon steel [14]. Microcracks were
also observed in specimens obtained from the failed shaft from regions close to the initial crack as shown in Fig
7(b).
160 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163

Fig. 7 Microstructure of the failed shaft

4.6 Fractography

The EDS spectra of the inclusion and the matrix are shown, respectively, in Figure 8. The inclusion spectrum
contains P peak while there is no such peak in the matrix spectrum. This means that there is higher P content in
the inclusion than in the matrix. Line scanning and elemental mapping of P revealed that P content is highest in
the area around the dark inclusion as shown in Figure 8. Other elements in the inclusion spectrum are Fe and C.
Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163 161

a) the dark inclusion and the crack along the fracture surface

b) EDS spectrum of an inclusion

Fig. 8 SEM fractograph of the failed shaft

162 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163

4.7 Stress analysis

The torque carried by individual shaft is 50% of the value in the equation 2 as there are two output shafts in the
gearbox as shown in Fig. 2. The values of Wp and T are 315 kW and 158 kN m. The values of W t and Wr are
363 and 132 kN, respectively.
The results of stress analysis showed that the maximum stress is 289.6 MPa as shown in Fig. 9.

Fig. 9 Maximum stress on wobbler corners of the shaft

4.8 Material fatigue strength

Fatigue strength or endurance limit for tempered steels can be approximated from tensile strength data using
equation 5 [9].

Se 0.50Sut (5)

e is endurance limit of material in MPa , Sut is tensile strength of shaft material in MPa.
For AISI 1045 steel, the tensile strength is 700 MPa [15]. The endurance limit of the steel is therefore
approximately 350 MPa. The endurance limit of the shaft is expected to be lower than the above figure due to
geometrical and other factors. The value can be calculated using the simplified equation 6 [9]

Se K a K b K c K d K e K f Se (6)

where Se is endurance limit of the shaft under investigation, ka is surface factor = 0.89, kb is size factor = 0.718,
kc is reliability factor = 0.897 for 90 % reliability, kd is temperature factor = 1 (assumed), ke is modifying factor
for stress concentration = 1, and kf is miscellaneous effects factor = 1. The endurance limit of this particular shaft,
calculated using the above data, is approximately 201 MPa.
Resuls from the stress analysis showed that the maximum stress is 289.6 MPa. From the stress at base on the
results from analysis, it was noticed that the maximum stress exceeding the yield stress is in some regions of the
wobblers surface . The stress at the wobblers surface of the shaft is 289.6 MPa, higher than the endurance limit of
the shaft material. The presence of the wobblers surface therefore cannot be, on its own, responsible for the
fatigue crack.
 Charnont Moolwan and Samroeng Netpu / Procedia - Social and Behavioral Sciences 88 (2013) 154 – 163 163

5. Discussions

The normal hardness was 290 HV and 210 240 HV for welding zone (A) and shaft material (B),
respectively.
The wear of wobbler of the shaft occurred at large dynamic loads during mill operation, particularly when the
piece being rolled was gripped and when it was rejected. These loads action to small contact area led to rapid
wear of the wobblers, an effect which intensified with an increase in the gaps between the coupling and the
wobblers.
The crack started at the welded region. This is because there is both a high residual stress and very high stress
concentration at this location.

6. Conclusions and recommendation

The failed shaft of the two high gearbox under this investigation failed by fatigue fracture.
The presence of inclusions in the structure is the root cause of the failure.
The fatigue crack has initiated because resulting from the weak zone gave rise to high stress concentration
leading to crack initiation and propagated inwards until final fracture occurred.
Proper preheat during welding could have prevented this failure. Proper preheat is necessary when AISI 1045
steel is welded.
Improve wear resistance of the wobblers of the shaft was thermal spraying for more hardness.

Acknowledgement

The author thanks Mr. Tawat Sopasit, the manager of TTS steel mill Co., Ltd for his help in providing
information about the history of the failed shaft and allowing the publication of this information.

References

[1] Available from: http:/www.elecon.com gearworld/dat-gw-failure.html

[2] Hamrock BJ, Schmind SR, Jacoson OB.Fundamentals of machine elements. Second Ed. New York: McGraw-Hill; 2005.
[3] Berndt F, Bennekom van A. Pump Shaft Failure-a Compendium of Case Studies. Engineering failure analysis, 2001
[4] Xu Xiaolei,Yu Zhiwei, Ding Hongxin. Failure Analysis of Diesel engine gear shaft. Engineering Failure Analysis, 2006
[5] Failure analysis and prevention ASM handbook, vol. 11. Metals Park (OH): American Society for Metals; 1996.
[6] S.C Rulmenti, URB ball and roller bearings catalogue
[7] Available from: http:/www.engineeringtoolbox.com/electrical-formulars-d_455.html
[8] Handbook of engineering fundamentals. Wiley Engineering Handbook Series, John Wiley&Son, NY, 1975. p.537-541
[9] Shigley JE. Mechanical Engineering Design. Metric Edition, New York: McGraw-Hill Book Company; 1986. p. 242-256
[10] Available from: http:/www.bssteel.co.th/product_en.html
[11] Available from: http:/www.efunda.com/materials/alloys/carbon_ steels
h ed.,
McGraw-Hill, 1987.
[13] S.H. Loewenthal, Shafts, Couplings, Keys, ETC, Mechanical Design Handbook, McGraw-Hill, New York.
[14] W.F Smith, Structure and Properties of Engineering Alloys, 2nded., McGraw-Hill Inc., Singapore 1993.
[15] Available from: http:/www.globalmetals.com.au